The Voltage Prediction of a Buck Converter Using Machine Learning Approaches

Machine learning is a part of artificial intelligence and a method of analyzing data that convert to automatic operations the building of analytical models. It is aimed on the proposal that systems can learn from materials on their own, identify patterns, and build decisions with little or no human assistance. The paper reviews the machine learning as a process for teaching computers to learn from experience or directly from data, based on a predefined equation as a model. We used four types of decision tree as machine learning methods for data set classification, such as PERTree, M5P, RandomTree and RandomForest. First, we give the equations for buck converter as a model, then we teach the computer to make predictions by his own. The Buck DC-DC converter decrease voltage by using a transformer, so the output voltage is always less than or equal to the input voltage. Second, the way we gain the database and WEKA software are described. WEKA operate with.arff file format, so we first convert our database in the required format. Then we present and discuss the results obtained using different types of classification.